[E-trademarks] Help on quick, down, and dirty learning the new advanced search features

[Neil R. Ormos]

Ken Boone via E-trademarks wrote:

> A new search theory: the search syntax
> GS3:/.*009.*widget.*/ is effective for
> retrieving all records with the single term
> widget in class 009.

> My immediate reaction: Where did the search
> field GS3 come from?  I recalled the GS2 search
> field but not the GS3 search field.  Checking
> the online help, following are the relevant list
> of search fields. [...]

> [...] Another example is 
> 
> GS3:/.*009.*INTEGRATED.CIRCUITS.*/ AND LD:true AND SN:72* 
> 
> for which I expected to see 72457899 among the
> hits. Nope. [...]

> That's 3 examples with specific serial numbers
> that were not included in the search results.
> So what am I doing wrong?  Or have I presented 3
> counter examples to that search theory?

Compare the results from

GS3:/.*009.*integrated.circuit.*/ AND LD:true AND SN:72*

with 

GS3:/.*021.*integrated.circuit.*/ AND LD:true AND SN:72*

which latter set does include 72457899.  

In the single result for the 009 search, "009" is contained in the G & S field.  

In the results for the 021 search, "021" is contained in the G & S fields, but "009" is not.  

Maybe we need to ask the PTO to add another derivative index, say "GS4", that, for each class, contains GS3, with the international class code prepended.

It would also be nice if the PTO would kindly clarify whether search terms between slashes are interpreted as anchored POSIX regular expressions.